CN112256590B - Virtual scene effectiveness judgment method and device and automatic driving system - Google Patents
Virtual scene effectiveness judgment method and device and automatic driving system Download PDFInfo
- Publication number
- CN112256590B CN112256590B CN202011263668.7A CN202011263668A CN112256590B CN 112256590 B CN112256590 B CN 112256590B CN 202011263668 A CN202011263668 A CN 202011263668A CN 112256590 B CN112256590 B CN 112256590B
- Authority
- CN
- China
- Prior art keywords
- virtual scene
- data
- driving risk
- simulation test
- test vehicle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
Abstract
The application relates to a virtual scene effectiveness judgment method and device, an automatic driving system and a storage medium. The method comprises the following steps: acquiring virtual scene driving risk data and virtual scene sending time delay of a simulation test vehicle in a virtual scene; correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain corrected driving risk data; acquiring the historical accident rate of a simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene; and judging the effectiveness of the virtual scene according to the matching result. The interference of the invalid virtual scene on the simulation test is avoided, so that the accuracy of the simulation test result is improved in the subsequent simulation test process.
Description
Technical Field
The application relates to the technical field of artificial intelligence, in particular to a virtual scene validity judgment method and device, an automatic driving system and a storage medium.
Background
With the development of artificial intelligence technology, automatic driving technology appears, and simulation testing is an important process in the automatic driving technology. During simulation test, the simulation test platform issues virtual scenes to the test vehicle, and the driving safety conditions of the scenes are reflected through the virtual scenes.
In the traditional simulation test technology, a test vehicle tests based on a received virtual scene, however, the virtual scene information is interfered by a plurality of factors in the issuing process, and a scheme for judging the validity of the virtual scene is not provided in the existing automatic driving simulation test, so that the virtual scene information received by the vehicle may not completely reflect the driving safety of the scenes, and the high-precision simulation test cannot be supported.
Disclosure of Invention
In view of the above, it is necessary to provide a virtual scene validity determination method, device, automatic driving system, and storage medium capable of improving accuracy of simulation test results in view of the above technical problems.
A virtual scene effectiveness judgment method comprises the following steps:
acquiring virtual scene driving risk data and virtual scene sending time delay of a simulation test vehicle in a virtual scene;
correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain corrected driving risk data;
acquiring the historical accident rate of a simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene;
and judging the effectiveness of the virtual scene according to the matching result.
A virtual scene validity determination apparatus, the apparatus comprising:
the data acquisition module is used for acquiring virtual scene driving risk data and virtual scene sending time delay of the simulation test vehicle in a virtual scene;
the data correction module is used for correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain corrected driving risk data;
the data matching module is used for acquiring the historical accident rate of the simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene;
and the validity judging module is used for judging the validity of the virtual scene according to the matching result.
An autopilot system comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of:
acquiring virtual scene driving risk data and virtual scene sending time delay of a simulation test vehicle in a virtual scene;
correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain corrected driving risk data;
acquiring the historical accident rate of a simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene;
and judging the effectiveness of the virtual scene according to the matching result.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring virtual scene driving risk data and virtual scene sending time delay of a simulation test vehicle in a virtual scene;
correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain corrected driving risk data;
acquiring the historical accident rate of a simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene;
and judging the effectiveness of the virtual scene according to the matching result.
According to the virtual scene validity judging method, the virtual scene validity judging device, the automatic driving system and the storage medium, the virtual scene driving risk data and the virtual scene sending time delay of the simulation test vehicle in the virtual scene are obtained, the time delay caused by sending the virtual scene is taken as a consideration factor, the driving risk in the virtual scene where the vehicle is located is corrected according to the sending time delay and the accident response time preset by the simulation test vehicle, the historical traffic accident rate is taken as the prior information of the virtual scene, the corrected driving risk data is matched with the historical accident rate, whether the virtual scene can reflect the real scene or not is judged according to the matching result, the validity of the virtual scene is determined, the interference of the invalid virtual scene on the simulation test is avoided, and therefore the accuracy of the simulation test result in the subsequent simulation test process is improved.
Drawings
FIG. 1 is a diagram of an application environment of a virtual scene validity determination method in an embodiment;
FIG. 2 is a flowchart illustrating a virtual scene validity determination method according to an embodiment;
FIG. 3 is a schematic flowchart of a virtual scene validity determination method in another embodiment;
FIG. 4 is a flowchart illustrating a virtual scene validity determination method according to yet another embodiment;
FIG. 5 is a flowchart illustrating a virtual scene validity determination method according to another embodiment;
FIG. 6 is a flowchart illustrating a virtual scene validity determination method according to yet another embodiment;
FIG. 7 is a flowchart illustrating a virtual scene validity determination method according to another embodiment;
FIG. 8 is a diagram of an application scenario of a virtual scenario validity determination method in one embodiment;
FIG. 9 is a diagram illustrating another exemplary embodiment of a virtual scene validity determination method;
FIG. 10 is a diagram illustrating another exemplary embodiment of a virtual scene validity determination apparatus;
FIG. 11 is a block diagram showing the configuration of a virtual scene validity determination apparatus according to an embodiment;
fig. 12 is an internal structural view of an automatic driving system in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Artificial Intelligence (AI) is a theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human Intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making.
The automatic driving technology is an important application aspect of artificial intelligence, generally comprises technologies such as high-precision maps, environment perception, behavior decision, path planning, motion control and the like, and has wide application prospects.
The virtual scene effectiveness judgment method provided by the application can be applied to the application environment shown in fig. 1. Wherein the simulation test vehicle 102 communicates with the simulation test platform 104 over a network. For example, by cellular base station 106, and the like. The simulation test platform 104 sends the virtual scene to the simulation test vehicle 102, and the simulation test platform 104 receives the virtual scene driving risk data and the virtual scene sending time delay fed back by the simulation test vehicle 102 according to the virtual scene, corrects the virtual scene driving risk data according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle, and obtains corrected driving risk data; acquiring the historical accident rate of a simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the actual accident scene corresponds to the virtual scene; and judging the effectiveness of the virtual scene according to the matching result. The simulation test platform 104 may be implemented by a server, and the server may be implemented by an independent server or a server cluster composed of a plurality of servers. In other embodiments, the data processing process of the simulation test platform may also be implemented by any other computer device having an interaction channel with the simulation test platform, which is not limited herein.
In an embodiment, as shown in fig. 2, a virtual scenario validity determination method is provided, which is described by taking the method as an example for being applied to the simulation test platform in fig. 1, and includes the following steps 202 to 208.
Step 202, acquiring virtual scene driving risk data and virtual scene transmission time delay of a simulation test vehicle in a virtual scene.
The virtual scene is a simulation scene generated by a simulation simulator in the simulation test platform through a simulation technology, the virtual scene is not a real scene, and various environmental information can be simulated through the simulation technology. Taking a virtual scene in the automatic driving process as an example, the virtual scene comprises data information of other traffic participants, such as quality, speed, acceleration, GPS coordinates, pavement viscosity, pavement friction coefficient, pavement curvature, visibility and the like.
The simulation test platform sends the virtual scene to the simulation test vehicle, and the vehicle-mounted computer of the simulation test vehicle analyzes the driving risk of the virtual scene by extracting the data information in the virtual scene and combining the self running parameters of the simulation test vehicle to obtain the driving risk data of the virtual scene. The virtual scene driving risk data refers to the probability of accidents of a simulation test vehicle in a virtual scene.
The virtual scene sending time delay refers to the time difference from the time point when the simulation test platform sends the virtual scene to the time point when the simulation test vehicle receives the virtual scene.
In an embodiment, the virtual scene sending delay may be determined by a vehicle-mounted computer of the simulation test vehicle according to sending time carried by the virtual scene and receiving time of the virtual scene, and the vehicle-mounted computer of the simulation test vehicle feeds back the determined virtual scene sending delay to the simulation test platform. Or the receiving time of the virtual scene can be fed back to the simulation test platform by a vehicle-mounted computer of the simulation test vehicle, and the simulation test platform determines the receiving time according to the pre-recorded sending time and feedback of the virtual scene.
In one embodiment, as shown in fig. 3, the step 202 of acquiring the virtual scene driving risk data and the virtual scene transmission delay of the simulation test vehicle in the virtual scene includes steps 302 to 304.
Step 302, sending virtual scene data carrying time data and environmental parameters to a simulation test vehicle.
Step 304 is to receive the virtual scene driving risk data and the virtual scene sending time delay fed back by the simulation test vehicle.
The virtual scene driving risk data is obtained by the simulation test vehicle according to the environmental parameters in the virtual scene and the operation parameters of the simulation test vehicle, and the virtual scene sending time delay is obtained by the simulation test vehicle according to the receiving time of the virtual scene and the time difference of the sending time.
The simulation test vehicle obtains the self quality, speed, acceleration and GPS coordinates from the in-vehicle equipment, extracts the quality, speed, acceleration, GPS coordinates, road viscosity, road friction coefficient, road curvature and visibility of other traffic participants in the virtual scene from the virtual scene issued by the simulation test platform, and then calculates the vehicle-end driving risk by adopting the existing driving risk calculation model.
In an embodiment, the calculation formula of the potential collision risk of the vehicle, namely the driving risk data of the virtual scene is as follows
Wherein, mass M, distance R, speed V, driving direction is cos theta, DR refers to driver risk factor, R refers to road surface factor, and the road surface factor includes viscosity, humidity and gradient of the road surface. And (3) temperature. D is the road width. LTa is the type of road sign, and takes values of 1,2, 3 and 4 (the larger the pressure brought to the driver by the road sign, the larger the value of LTa). k is a radical of1＝3，k2＝1，k3The speed of light.
The objects corresponding to other traffic participants are divided into two categories: the stationary object is classified into two types, namely a stationary object which is stationary but collides with the host vehicle, and a stationary object which is stationary and does not collide with the host vehicle but affects the collision of the host vehicle with other objects. The stationary and moving are calculated using the first and second equations, respectively.
The first term of the first formula is a calculation method of the collision strength between the host vehicle and an object that is stationary but is likely to collide with the host vehicle. The second term of the first formula is a calculation method of the driving risk of the host vehicle caused by a stationary object that does not collide with the host vehicle but affects the driving risk of the host vehicle. The second formula is a calculation method of a driving risk between an object that is moving and may collide with the host vehicle and the host vehicle, and the driving risk may refer to a collision probability between vehicles or a collision strength between vehicles.
And analyzing and calculating to obtain the driving risk data of the virtual scene by the vehicle-mounted computer of the simulation test vehicle based on the data information in the virtual scene and by combining the self running parameters of the simulation test vehicle.
In one embodiment, when the simulation test platform sends the virtual scene, a timestamp corresponding to sending time is marked on the virtual scene, when the vehicle-mounted computer of the simulation test vehicle receives the virtual scene, a timestamp corresponding to receiving time is marked on the virtual scene, and the sending time delay of the virtual scene can be obtained according to the time difference of the two timestamps corresponding to the virtual scene.
After the vehicle-mounted computer of the simulation test vehicle obtains the virtual scene driving risk data and the virtual scene sending delay corresponding to the virtual scene through analysis, the virtual scene driving risk data and the virtual scene sending delay are fed back to the simulation test platform through the network, and the simulation test platform can obtain the virtual scene driving risk data and the virtual scene sending delay through receiving the feedback data. The analysis process of the driving risk data of the virtual scene and the transmission time delay of the virtual scene is realized by concentrating on a vehicle-mounted computer of a simulation test vehicle, so that the data transmission quantity can be reduced, the occupation of data transmission resources is reduced, fewer data transmission processes are adopted, the data result can be obtained more quickly, and the data processing efficiency is improved.
And 204, correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle to obtain the corrected driving risk data.
The accident response time is a time for responding to a situation occurring in a virtual scene, and due to the existence of a delay, the real-time performance of a driving risk is damaged. However, the driver or the control system in the simulation test vehicle (the controller of the autonomous vehicle is the control system) has a certain tolerance to the disruption of real-time performance, which is related to the accident response time of the driver or the control system. The reaction time of the vehicle driver or the control system is preset. For example, the reaction time of the driver is set to 0.7 second or the like.
In the embodiment, the correction parameter of the virtual scene driving risk data can be determined according to the value of the virtual scene sending time delay and the accident response time preset by the simulation test vehicle, the correction parameter can be specifically the ratio result of the smaller value and the larger value in the virtual scene sending time delay and the accident response time preset by the simulation test vehicle, the virtual scene driving risk data is corrected according to the correction parameter, the corrected driving risk data is obtained, and therefore the driving risk data in the virtual scene where the simulation test vehicle is located can be corrected according to the time delay in the virtual scene sending process, and more accurate and reliable driving risk data can be obtained.
And step 206, acquiring the historical accident rate of the simulation test vehicle in the actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result.
Wherein the actual accident scenario corresponds to the virtual scenario. The actual accident scene is a scene which has actually occurred, and the actual accident scene corresponds to a virtual scene, and the difference is that the virtual scene is generated by a virtual structure, and the actual accident scene is a history scene which actually occurs.
The simulation test platform has historical accident rate in the corresponding actual accident scene for each type of virtual scene of the place of origin, the corresponding actual accident scene can be determined according to the type of the virtual scene, and then the historical accident rate of the simulation test vehicle in the actual accident scene is obtained through searching the actual accident scene.
The historical traffic accident rate of a certain traffic accident scene reflects the probability that the simulation test vehicle in the virtual scene encounters traffic accident risks in the virtual scene, and for a certain simulation test vehicle, the corrected driving risk data of the certain simulation test vehicle is in a forward relation with the traffic accident risks, namely the historical accident rate, of the actual accident scene. And judging whether the virtual scene sent by the simulation test platform to the simulation test vehicle is effective or not according to whether the matching result is in a forward relation or not by matching the corrected driving risk data with the historical accident rate.
And step 208, judging the effectiveness of the virtual scene according to the matching result.
In the embodiment, whether the virtual scene sent by the simulation test platform to the simulation test vehicle is effective or not is judged according to whether the matching result of the corrected driving risk data and the historical accident rate is in a forward relation or not. Specifically, when the matching result of the corrected driving risk data and the historical accident rate is in a positive relationship, the virtual scene issued by the simulation test platform to the simulation test vehicle is judged to be valid, and when the matching result of the corrected driving risk data and the historical accident rate is not in the positive relationship, the virtual scene issued by the simulation test platform to the simulation test vehicle is judged to be invalid.
According to the virtual scene effectiveness judging method, the virtual scene driving risk data and the virtual scene sending time delay of the simulation test vehicle in the virtual scene are obtained, the time delay caused by virtual scene sending is taken as a consideration factor, the driving risk in the virtual scene where the vehicle is located is corrected according to the sending time delay and the accident response time preset by the simulation test vehicle, the historical traffic accident rate is taken as the prior information of the virtual scene, whether the virtual scene can reflect the real scene or not is judged by matching the corrected driving risk data with the historical accident rate, the effectiveness of the virtual scene is determined, the interference of the invalid virtual scene on the simulation test is avoided, and therefore the accuracy of the simulation test result in the subsequent simulation test process is improved.
In one embodiment, as shown in fig. 4, the driving risk data of the virtual scene is modified according to the virtual scene transmission delay and the accident response time preset by the simulation test vehicle, so as to obtain modified driving risk data, that is, step 204, which includes steps 402 to 404.
And 402, when the virtual scene sending time delay is smaller than the accident response time, correcting the virtual scene driving risk data according to the ratio of the virtual scene sending time delay to the accident response time to obtain corrected driving risk data.
And 404, when the virtual scene sending time delay is not less than the accident response time, correcting the virtual scene driving risk data according to the ratio of the accident response time to the virtual scene sending time delay to obtain corrected driving risk data.
Since the reaction of the simulated test vehicle to the driving risk of the virtual scene is only useful at the moment the virtual scene is generated, the delay t if the simulated test vehicle i receives the virtual scene ji,jTime delay T of driver or control systemiEven if the simulation test vehicle reacts to the driving risk, the actual meaning of the reaction is very small, the positive influence on the traffic accident rate is small, and the negative influence is large; if the simulation test vehicle i receives the delay t of the virtual scene ji,jLess than TiThen, ti,jThe closer to TiMeaning that the less meaningful this driving risk data is, the greater the negative impact on the traffic accident rate, whereas the more meaningful is, the less negative impact on the traffic accident rate. Thus, i.e. for the simulation test vehicle i, the delay t when the simulation test vehicle i receives the virtual scene ji,jTime delay T of driver or control systemiIts driving risk E in the virtual scene ji,jShould be modified to
Time delay t of receiving virtual scene j by simulation test vehicle ii,jGreater than or equal to TiDriving risk E in virtual scene ji,jShould be modified to
The driving risk data in the virtual scene where the simulation test vehicle is located is corrected according to the delay of the virtual scene sending process, so that more accurate and reliable driving risk data are obtained.
In one embodiment, as shown in fig. 5, the number of virtual scenes is plural. Acquiring the historical accident rate of the simulation test vehicle in the actual accident scene, and matching the corrected driving risk data with the historical accident rate to obtain a matching result, wherein the step 206 comprises steps 502 to 506.
Step 502, obtaining historical accident rates of the simulation test vehicle in actual accident scenes corresponding to the virtual scenes, and sequencing the corrected driving risk data and the historical accident rates corresponding to the virtual scenes of the simulation test vehicle according to the numerical values.
And step 504, constructing a data pair according to the corrected driving risk data and the historical accident rate which have the same sequencing order.
And step 506, when the corrected driving risk number in the data pair is different from the virtual scene corresponding to the historical accident rate, marking the data pair as a matching failure data pair.
For the simulation test vehicle i, when the simulation test vehicle i receives the delay t of the virtual scene ji,jTime delay T of driver or control systemiIts driving risk E in the virtual scene 1,2i,1,Ei,2,...,Ei,nShould be respectively modified into
Time delay t of receiving virtual scene j by simulation test vehicle ii,jGreater than or equal to TiThe driving risk E in the virtual scene 1,2i,1,Ei,2,...,Ei,nShould be respectively modified into
Modified driving risk data, respectively
For the simulation test vehicle i, the corrected driving risk data in the virtual scenes 1,2
Because the historical traffic accident rate of a certain traffic accident scene reflects the probability that the simulation test vehicle in the virtual scene encounters the traffic accident risk in the virtual scene, the corrected driving risk data should have a forward relationship with the traffic accident risk of the actual accident scene, namely, the historical traffic accident rate p1,p2,...,pnShould be correlated with the corrected driving risk data
The magnitude relationship between them is the same. The simulation test platform judges whether the virtual scene received by the simulation test vehicle can reflect the traffic accident situation of the real scene according to the following method:
p is to be1,p2,...,pnAnd
the numerical values are arranged from large to small or from small to large in sequence, and the numerical values are recorded after arrangement
And
then comparing the same sequence numbers
And
see if they correspond to the same scene, if not, they are marked as "matching failed data pairs", otherwise, they are called "matching successful data pairs".
By arranging the values and constructing the data pair form, whether the corrected driving risk data and the historical accident rate are in a forward relation or not can be quickly detected. Thereby increasing data processing speed.
In one embodiment, determining the validity of the virtual scene based on the matching result comprises: and obtaining ratio data of the number of the matching failure data to the number of the data pairs according to the number of the data pairs corresponding to the simulation test vehicle and the number of the matching failure data pairs. And when the ratio data is smaller than the historical accident rate, judging that the virtual scene is effective. And when the ratio data is not less than the historical accident rate, judging that the virtual scene is invalid.
And comparing the number of all the 'matching failure data pairs' with the number of all the 'data pairs' to obtain the proportion of the 'matching failure data pairs'. And judging whether the traffic accident rate is smaller than the historical accident rate, wherein one of the indexes of the actual traffic is to reduce the traffic accident rate because the simulation test result is used for guiding the actual traffic, and one of the indexes of the actual traffic is to reduce the traffic accident rate, so that the test result is beneficial to reducing the traffic accident rate of the area, namely the average traffic accident rate of various scenes. Therefore, if the proportion of the matching failure data pair is greater than the historical accident rate of the corresponding virtual scene, the final simulation result is not beneficial to reducing the traffic accident rate, so that the actual traffic is not guided to reduce the traffic accident rate, and if the proportion of the matching failure data pair is less than the historical accident rate of the corresponding virtual scene, the virtual scene issued by the simulation test platform can reflect the traffic accident situation of the real scene; if the former is larger than the latter, the simulation can not reflect the traffic accident situation of the real scene.
In one embodiment, the number of simulation test vehicles is plural. As shown in fig. 6, determining the validity of the virtual scene according to the matching result includes steps 602 to 608.
Step 602, aggregating the number of data pairs corresponding to each simulation test vehicle and the number of matching failure data pairs to obtain the total number of data pairs and the total number of matching failure data pairs.
Step 604, determining ratio data of the total amount of the matching failure data pairs and the total amount of the data pairs, and determining an average value of historical accident rates of the simulation test vehicles.
And step 606, judging that the virtual scene is effective when the ratio data is smaller than the average value.
And step 608, when the ratio data is not less than the average value, judging that the virtual scene is invalid.
For the simulation test vehicles 1, 2.. and m, "matching failure data pairs" of all the simulation test vehicles are determined. The number of "matching failure data pairs" corresponding to the vehicles 1, 2.. multidot.m is recorded as u1,u2,...,um(ii) a Comparing all the matching failure data pairs with all the data pairs to obtain matching lossesThe ratio of the failure data to the occupation ratio. All "matching failure data pairs" are u in number1+u2+...+umAll vehicles 1,2,.. m have the same number n "data pairs" all "data pairs" number nm, their ratio being
And determining whether it is less than the average historical accident rate
If the ratio data is smaller than the average historical accident rate, the virtual scene issued by the simulation test platform can reflect the traffic accident situation of the real scene; if the ratio data is larger than or equal to the average historical accident rate, the simulation can not reflect the traffic accident situation of the real scene.
In one embodiment, the method further comprises: and when the virtual scene is judged to be invalid, adjusting the transmission time delay of the virtual scene or the accident response time preset by the simulation test vehicle to obtain the effective virtual scene.
For the simulation test vehicle i, the historical traffic accident rate p of the simulation test vehicle i in traffic accident scenes 1,21,p2,...,pnAnd correcting driving risk data
Arranged according to numerical values from large to small or from small to large, respectively
And
then compare the rank-identical
And
whether or not they correspond to the same scene, and if not, they are said to be associated with a simulationThe "matching failure data pair" corresponding to the test vehicle i. All "matching failure data pairs" corresponding to the simulation test vehicle i are found. Because, for the "matching failure data pair" corresponding to the simulation test vehicle i "
And
suppose and
the corresponding traffic accident scenario is scenario ykAt this time, it is not necessary to consider
Corresponding virtual scene, there must be another matching failure data pair, which is recorded as
And
suppose and
the corresponding traffic accident scenario is yzAt this time, it is not necessary to consider
Corresponding virtual scene, therefore, when the virtual scene is issued next, the simulation test vehicle i should be in the traffic accident scene ykThe data of the lower encountered corrected driving risk are controlled in
Within.
The modified driving risk data control method includes: and adjusting the issuing delay of the virtual scene information and the reaction time of a vehicle driver or a control system. For example, adjusting the reception of vehicle i into virtual scene ykIs delayed by
Or the reaction time T of the vehicle control systemiTo make
And
corresponding to the same traffic accident scene, placing the vehicle i in the traffic accident scene yzThe data of the lower encountered corrected driving risk are controlled in
To inside so that
And
corresponding to the same traffic accident scene, the aim is to make the matching failure data pair u1+u2+...+umIn comparison with all "data to" quantities being nm
Controlling average historical traffic accident rate
And adjusting the invalid virtual scene into the valid virtual scene.
In one embodiment, as shown in fig. 7, a virtual scene validity determination method is provided, which includes the following steps:
step 702, sending the virtual scene with the sending time to the simulation test vehicle, so that the simulation test vehicle obtains the virtual scene sending time delay according to the receiving time and the sending time of the virtual scene, and obtains the virtual scene driving risk data according to the environment parameters in the virtual scene and the operation parameters of the simulation test vehicle.
Step 704, sending the virtual scene data carrying the time data and the environmental parameters to the simulation test vehicle.
Step 704 receives the virtual scene driving risk data and the virtual scene sending time delay fed back by the simulation test vehicle.
And 708, when the virtual scene transmission delay is not less than the accident response time, correcting the virtual scene driving risk data according to the ratio of the accident response time to the virtual scene transmission delay to obtain corrected driving risk data.
Step 710, obtaining a historical accident rate of the simulation test vehicle in an actual accident scene, wherein the actual accident scene corresponds to the virtual scene.
And 712, sequencing the corrected driving risk data and the historical accident rate corresponding to each virtual scene of the simulation test vehicle according to the numerical values.
And 714, constructing a data pair according to the corrected driving risk data and the historical accident rate with the same sequencing sequence.
And step 716, when the corrected driving risk number in the data pair is different from the virtual scene corresponding to the historical accident rate, marking the data pair as a matching failure data pair.
Step 718, aggregating the number of data pairs corresponding to each simulation test vehicle and the number of matching failure data pairs to obtain the total number of data pairs and the total number of matching failure data pairs.
And 720, determining ratio data of the total amount of the matching failure data pairs and the total amount of the data pairs, and determining the average value of the historical accident rate of each simulation test vehicle.
And step 722, judging that the virtual scene is effective when the ratio data is smaller than the average value.
And step 724, when the ratio data is not less than the average value, judging that the virtual scene is invalid.
And 726, adjusting the virtual scene sending time delay or the accident response time preset by the simulation test vehicle to obtain an effective virtual scene.
As shown in fig. 8, the present application further provides an application scenario, where the virtual scenario validity determination method can be applied to a platform in a vehicle test field, and the application scenario includes vehicle in-loop simulation, that is, scenario information is issued to a test vehicle through a simulation test platform for vehicle testing. The application scenario can include testing real vehicles, virtual pedestrians, virtual non-motor vehicles and the like. The application scene applies the virtual scene validity judgment method. Specifically, the virtual scene validity determination method is applied to the application scene as follows:
the vehicle determines the issuing delay of the scene information: the simulation test platform simulates a plurality of historical traffic accident scenes, the number of the scenes is recorded as n, the n scenes are respectively called as scenes 1,2, the1,p2,...,pn. Each scene information contains a timestamp; the simulation test platform issues the scenes to the vehicles, and each vehicle receives the scenes 1, 2. The vehicle stamps the scene information when receiving each scene information, so that the issuing delay of each scene information can be determined. Recording that m vehicles are in the target area, namely vehicles 1,2,. and m; recording the time delay of the vehicle i receiving the virtual scene j as ti,j(ii) a After receiving a scene issued by the simulation test platform, the vehicle determines the driving risk in the issued scene: the vehicle obtains the self mass, speed, acceleration and GPS coordinates from the in-vehicle equipment, extracts the mass, speed, acceleration, GPS coordinates, road surface viscosity, road surface friction coefficient, road curvature and visibility of other traffic participants in the scene from the virtual scene issued by the test platform, and then calculates the vehicle end driving risk by adopting the existing driving risk calculation model. Recording the driving risk of the vehicle i in the virtual scene k as Ei,k(ii) a And the vehicle corrects the driving risk according to the time delay: because of the existence of the time delay, the real-time property of the driving risk is damaged. However, the driver or the control system has a certain tolerance to the disruption of the real-time behavior, which is related to the reaction time of the driver. Note that the reaction time of the driver or control system of the vehicle 1,21,T2,...,Tm(vehicle driver or control systemThe system delay is predetermined and known. For example, the reaction time of the driver is 0.7 seconds); since the vehicle's reaction to the driving risk of the scene is useful at the moment the scene is generated, if the vehicle i receives the delay t of the scene ji,jTime delay T of driver or control systemiEven if the vehicle reacts to the driving risk, the actual meaning of the reaction is very small, the positive influence on the traffic accident rate is small, and the negative influence is large; if vehicle i receives delay t of scene ji,jLess than TiThen, ti,jThe closer to TiMeaning that the greater the negative impact of this risk on the traffic accident rate, and vice versa, the smaller the negative impact on the traffic accident rate. The significance of the driving risk is therefore directly inversely proportional to the delay and inversely proportional to the reaction time of the driver or the control system, i.e. for vehicle i, its driving risk E in the scene 1,2i,1,Ei,2,...,Ei,nShould be respectively modified into
Obtaining corrected driving risk data, respectively
For vehicle i, its modified driving risk data in simulated traffic accident scenarios 1,2
Since the historical traffic accident rate of a certain traffic accident scene reflects the probability that a vehicle in the scene encounters a traffic accident risk in the scene, the corrected driving risk data should have a positive relationship with the traffic accident risk of the actual scene, i.e., the historical traffic accident rate p1,p2,...,pnShould have a magnitude relationship with significant driving risk
The magnitude relationship between them is the same. However, due to the influence of factors such as time delay, the tolerance to the difference is also certain. Thus, according toThe following method judges whether the virtual scene received by the vehicle can reflect the traffic accident situation of the real scene: p is to be1,p2,...,pnAnd
arranged from large to small or from small to large, respectively
And
then the
And
whether the data correspond to the same scene or not is checked, if not, the data are marked as matching failure data pairs, otherwise, the data are called as matching success data pairs; by analogy, all "matching failure data pairs" are determined for vehicles 1, 2. The number of "matching failure data pairs" corresponding to the vehicles 1, 2.. multidot.m is recorded as u1,u2,...,um(ii) a And comparing all the matching failure data pairs with all the data pairs to obtain the matching failure data pair. All "matching failure data pairs" are u in number1+u2+...+umThe vehicles 1,2, are all of the same data quantity n "data pairs", the total "data pair" quantity being nm, their ratio being
And determining whether it is less than the average historical traffic accident rate
If the traffic accident situation is smaller than the preset traffic accident situation, the virtual scene issued by the simulation test platform can reflect the traffic accident situation of the real scene, otherwise, the traffic accident situation of the real scene cannot be reflected.
In order to prove the reliability of the virtual scene validity judgment method in the application, the applicant performs a plurality of experiments. The experiment is carried out in a simulator, a virtual scene is randomly issued to a vehicle in the simulator through a simulation test platform in the simulator, and an experiment group and a comparison group are set for carrying out comparison experiment.
Control group: as long as the virtual scene sent by the simulation test platform exists in the database, the virtual scene is considered to be effective.
Experimental groups: the vehicle returns the risk calculation value to the simulation test platform, and the simulation test platform judges whether the virtual scene is effective or not.
The number of invalid virtual scenes which are mistakenly regarded as valid virtual scenes in the experimental group and the comparison group is respectively recorded, and the ratio of the number of the invalid virtual scenes which are mistakenly regarded as valid virtual scenes in the prior art and the scheme of the application is calculated. The results of the experiment are shown in table 1:
TABLE 1 results of the experiment
| Simulating fault conditions | Ratio of the number of false positives in the prior art and in the present application regarding invalid virtual scenes as valid |
| First experiment | 1.42 |
| Second experiment | 1.35 |
| Third experiment | 1.45 |
| Fourth experiment | 1.46 |
| Fifth experiment | 1.45 |
| The sixth experiment | 1.51 |
| The seventh experiment | 1.55 |
| The eighth experiment | 1.45 |
| The ninth experiment | 1.56 |
| The tenth experiment | 1.62 |
According to the experimental results, compared with the prior art, the scheme of the application can more accurately identify whether the virtual scene is the effective scene, and the reliability of the judgment result is higher.
The application further provides an application scene, and the virtual scene judged to be the effective scene by the virtual scene effectiveness judging method can be applied to the application scene. Specifically, the implementation environment of the application scenario is as shown in fig. 9, where the image display interface in fig. 9 includes a real road surface scene acquired by a real vehicle cockpit through a vehicle-mounted image or video acquisition device, a virtual scene at a first viewing angle generated by a simulator based on ground or aerial images or acquired data of the video acquisition device, and a virtual scene at a second viewing angle. The first viewing angle is the same as or similar to the viewing angle of the cockpit in the middle part of fig. 9, and the second viewing angle is the bird's eye viewing angle capable of showing the condition of the whole road vehicle in the lower right corner of fig. 9.
In addition, the application scenario shown in fig. 10 is further provided, where the scenario includes four scenario diagrams, a three-dimensional scenario diagram of an environment where the vehicle is located is shown in an upper left corner of the scenario diagram, a lane scenario diagram of the environment where the vehicle is located is shown in an upper right corner of the scenario diagram, a road test video acquired by the vehicle-mounted camera is shown in a lower right corner of the scenario diagram, and a simulation video generated by simulation based on data acquired by the camera fixed on the ground or in the air is shown in a lower left corner of the scenario diagram. The digital twinning is a simulation process integrating multidisciplinary, multi-physical quantity, multi-scale and multi-probability by fully utilizing data such as a physical model, sensor updating, operation history and the like, and realizes urban traffic flow simulation through fusion support of a high-precision map and industrial data to obtain a lane scene graph. In the application scene, parallel simulation and rendering acceleration can be performed through a cloud end, so that the relative precision of the simulated video obtained through simulation and the drive test video reaches 20cm and basically reaches the same precision. In addition, based on the feedback of the vehicle-mounted intelligent terminal, the effectiveness judgment of the simulation scene in the simulation video is realized, the effective real-time updating of the simulation video is realized, and the driving safety is improved.
For example, in the case that an accurate image cannot be acquired by the vehicle-mounted camera device in foggy weather or blocked turning sight, the current road condition information may not be truly reflected on the basis of the road test video acquired by the vehicle-mounted camera device, and the current road condition information can be more accurately reflected on the basis of the simulation of the data acquired by the ground or aerial camera device on the basis of the three-dimensional scene graph, the lane scene graph and the simulation video acquired by the simulation, which are generated based on the simulation of the data acquired by the ground or aerial camera device, without depending on the viewing angle of the vehicle, so that safety accidents are avoided. It should be understood that although the various steps in the flow charts of fig. 2-7 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-7 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
In one embodiment, as shown in fig. 11, there is provided a virtual scene validity determination apparatus 1100, which may be a software module or a hardware module, or a combination of the two modules, as a part of an automatic driving system, and specifically includes: a data acquisition module 1110, a data modification module 1120, a data matching module 1130, and a validity determination module 1140, wherein:
the data obtaining module 1110 is configured to obtain virtual scene driving risk data and virtual scene sending time delay of the simulation test vehicle in the virtual scene.
And the data correction module 1120 is configured to correct the driving risk data of the virtual scene according to the virtual scene transmission delay and the accident response time preset by the simulation test vehicle, so as to obtain corrected driving risk data.
The data matching module 1130 is configured to obtain a historical accident rate of the simulation test vehicle in an actual accident scene, and match the corrected driving risk data with the historical accident rate to obtain a matching result, where the actual accident scene corresponds to the virtual scene.
And an effectiveness judging module 1140, configured to judge effectiveness of the virtual scene according to the matching result.
In one embodiment, the data acquisition module is further configured to send the virtual scene with the sending time to the simulation test vehicle, so that the simulation test vehicle obtains the virtual scene sending time delay according to the receiving time and the sending time of the virtual scene, and obtains the virtual scene driving risk data according to the environmental parameters in the virtual scene and the operating parameters of the simulation test vehicle; and receiving virtual scene driving risk data and virtual scene sending time delay fed back by the simulation test vehicle.
In one embodiment, the data correction module is further configured to correct the driving risk data of the virtual scene according to a ratio of the virtual scene transmission delay to the accident reaction time when the virtual scene transmission delay is smaller than the accident reaction time, so as to obtain corrected driving risk data; and when the virtual scene sending time delay is not less than the accident response time, correcting the virtual scene driving risk data according to the ratio of the accident response time to the virtual scene sending time delay to obtain corrected driving risk data.
In one embodiment, the number of virtual scenes is multiple; the data matching module is also used for sorting the corrected driving risk data and the historical accident rate corresponding to each virtual scene of the simulation test vehicle according to the numerical value; constructing a data pair according to the corrected driving risk data and the historical accident rate which have the same sequencing sequence; and when the corrected driving risk number in the data pair is different from the virtual scene corresponding to the historical accident rate, marking the data pair as a matching failure data pair.
In one embodiment, the match results include a match failure data pair; the validity judging module is further used for obtaining ratio data of the number of the failed matching data to the number of the data pairs according to the number of the data pairs corresponding to the simulation test vehicle and the number of the failed matching data; when the ratio data is smaller than the historical accident rate, judging that the virtual scene is effective; and when the ratio data is not less than the historical accident rate, judging that the virtual scene is invalid.
In one embodiment, the match results include a match failure data pair; the number of the simulation test vehicles is multiple; the validity judging module is further used for collecting the number of the data pairs corresponding to each simulation test vehicle and the number of the matching failure data pairs to obtain the total number of the data pairs and the total number of the matching failure data pairs; determining ratio data of the total amount of the matching failure data pairs and the total amount of the data pairs, and determining an average value of historical accident rates of all simulation test vehicles; when the ratio data is smaller than the average value, judging that the virtual scene is effective; and when the ratio data is not less than the average value, judging that the virtual scene is invalid.
In one embodiment, the virtual scene validity determination apparatus further includes a data adjustment module, configured to adjust a virtual scene transmission delay or an accident response time preset by the simulation test vehicle, so as to obtain a valid virtual scene.
According to the virtual scene effectiveness judging device, the virtual scene driving risk data and the virtual scene sending time delay of the simulation test vehicle in the virtual scene are obtained, the time delay caused by the virtual scene sending is taken as a consideration factor, the driving risk in the virtual scene where the vehicle is located is corrected according to the sending time delay and the accident response time preset by the simulation test vehicle, the historical traffic accident rate is taken as the prior information of the virtual scene, whether the virtual scene can reflect the real scene or not is judged by matching the corrected driving risk data with the historical accident rate, the effectiveness of the virtual scene is determined, the interference of the invalid virtual scene on the simulation test is avoided, and therefore the accuracy of the simulation test result in the subsequent simulation test process is improved.
For specific limitations of the virtual scene validity determination apparatus, reference may be made to the above limitations of the virtual scene validity determination method, which is not described herein again. All or part of the modules in the virtual scene validity determination device may be implemented by software, hardware, or a combination thereof. The modules can be embedded in a hardware form or independent of a processor in the automatic driving system, and can also be stored in a memory in the automatic driving system in a software form, so that the processor can call and execute the corresponding operations of the modules.
In one embodiment, an autopilot system is provided that may include a server, the internal structure of which may be as shown in fig. 12. The autopilot system includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the autopilot system is configured to provide computational and control capabilities. The memory of the automatic driving system comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the automatic driving system is used for storing virtual scene effectiveness judgment data. The network interface of the automatic driving system is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a virtual scene validity determination method.
Those skilled in the art will appreciate that the configuration shown in fig. 12 is a block diagram of only a portion of the configuration associated with the present application and does not constitute a limitation on the autopilot system to which the present application is applied, and that a particular autopilot system may include more or fewer components than shown, or some components may be combined, or have a different arrangement of components.
In one embodiment, there is also provided an autopilot system comprising a memory in which is stored a computer program and a processor which, when executed by the processor, carries out the steps of the above-described method embodiments.
In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, in which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
In one embodiment, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The computer instructions are read from the computer-readable storage medium by a processor of the autopilot system, and the computer instructions are executed by the processor to cause the autopilot system to perform the steps in the method embodiments described above.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware related to instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (16)
1. A virtual scene validity determination method is characterized by comprising the following steps:
acquiring virtual scene driving risk data and virtual scene sending time delay of a simulation test vehicle in a virtual scene;
correcting the virtual scene driving risk data according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle and the ratio result of the smaller value and the larger value in the virtual scene sending time delay and the accident response time to obtain corrected driving risk data;
acquiring the historical accident rate of the simulation test vehicle in an actual accident scene, and matching the corrected driving risk data with the historical accident rate to determine whether the magnitude relation of the corrected driving risk data is the same as the magnitude relation of the historical accident rate, so as to obtain a matching result of whether the corrected driving risk data and the historical accident rate are in a positive relation, wherein the actual accident scene corresponds to the virtual scene;
and judging whether the virtual scene is effective or not according to whether the matching result is in a forward relation or not.
2. The method of claim 1, wherein the obtaining of the virtual scene driving risk data and the virtual scene transmission delay of the simulation test vehicle in the virtual scene comprises:
sending virtual scene data carrying time data and environmental parameters to a simulation test vehicle;
and receiving virtual scene driving risk data and virtual scene sending time delay fed back by the simulation test vehicle, wherein the virtual scene driving risk data is obtained by the simulation test vehicle according to environmental parameters in the virtual scene and operation parameters of the simulation test vehicle, and the virtual scene sending time delay is obtained by the simulation test vehicle according to the time difference between the receiving time and the sending time of the virtual scene.
3. The method according to claim 1, wherein the modifying the driving risk data of the virtual scene according to the virtual scene transmission delay and the accident response time preset by the simulation test vehicle and according to the ratio result of the smaller value and the larger value of the virtual scene transmission delay and the accident response time to obtain the modified driving risk data comprises:
when the virtual scene sending time delay is smaller than the accident response time, correcting the virtual scene driving risk data according to the ratio of the virtual scene sending time delay to the accident response time to obtain corrected driving risk data;
and when the virtual scene sending time delay is not less than the accident response time, correcting the virtual scene driving risk data according to the ratio of the accident response time to the virtual scene sending time delay to obtain corrected driving risk data.
4. The method of claim 1, wherein the number of virtual scenes is plural;
the matching the revised driving risk data to the historical accident rate comprises:
sorting the corrected driving risk data and the historical accident rate corresponding to each virtual scene of the simulation test vehicle according to the numerical value;
constructing a data pair according to the corrected driving risk data and the historical accident rate which have the same sequencing sequence;
and when the corrected driving risk number in the data pair is different from the virtual scene corresponding to the historical accident rate, marking the data pair as a matching failure data pair.
5. The method of claim 4, wherein determining whether the virtual scene is valid according to whether the matching result is in a forward relationship comprises:
obtaining ratio data of the number of the matching failure data to the number of the data pairs according to the number of the data pairs corresponding to the simulation test vehicle and the number of the matching failure data pairs;
when the ratio data is smaller than the historical accident rate, the matching result is that the corrected driving risk data and the historical accident rate are in a forward relation, and the virtual scene is judged to be effective;
and when the ratio data is not less than the historical accident rate, judging that the virtual scene is invalid if the corrected driving risk data and the historical accident rate are not in a forward relation according to the matching result.
6. The method of claim 4, wherein the number of simulated test vehicles is plural;
the determining whether the virtual scene is valid according to whether the matching result is in a forward relationship includes:
collecting the number of data pairs corresponding to each simulation test vehicle and the number of matching failure data pairs to obtain the total amount of the data pairs and the total amount of the matching failure data pairs;
determining ratio data of the total amount of the matching failure data pairs and the total amount of the data pairs, and determining an average value of historical accident rates of all simulation test vehicles;
when the ratio data is smaller than the average value, the matching result is that the corrected driving risk data and the historical accident rate are in a positive relation, and the virtual scene is judged to be effective;
and when the ratio data is not smaller than the average value, judging that the virtual scene is invalid if the matching result indicates that the corrected driving risk data and the historical accident rate are not in a forward relation.
7. The method of claim 5 or 6, further comprising:
and when the virtual scene is judged to be invalid, adjusting the transmission time delay of the virtual scene or the accident response time preset by the simulation test vehicle to obtain the effective virtual scene.
8. An apparatus for determining validity of a virtual scene, the apparatus comprising:
the data acquisition module is used for acquiring virtual scene driving risk data and virtual scene sending time delay of the simulation test vehicle in a virtual scene;
the data correction module is used for correcting the driving risk data of the virtual scene according to the virtual scene sending time delay and the accident response time preset by the simulation test vehicle and according to the ratio result of the smaller value and the larger value in the virtual scene sending time delay and the accident response time to obtain corrected driving risk data;
the data matching module is used for acquiring the historical accident rate of the simulation test vehicle in an actual accident scene, matching the corrected driving risk data with the historical accident rate to determine whether the magnitude relation of the corrected driving risk data is the same as the data magnitude relation of the historical accident rate, and obtaining a matching result whether the corrected driving risk data and the historical accident rate are in a positive relation, wherein the actual accident scene corresponds to the virtual scene;
and the validity judging module is used for judging whether the virtual scene is valid or not according to whether the matching result is in a forward relation or not.
9. The device of claim 8, wherein the data acquisition module is further configured to send virtual scene data carrying time data and environmental parameters to the simulation test vehicle; and receiving virtual scene driving risk data and virtual scene sending time delay fed back by the simulation test vehicle, wherein the virtual scene driving risk data is obtained by the simulation test vehicle according to environmental parameters in the virtual scene and operation parameters of the simulation test vehicle, and the virtual scene sending time delay is obtained by the simulation test vehicle according to the receiving time of the virtual scene and the time difference of the sending time.
10. The device according to claim 8, wherein the data modification module is further configured to modify the virtual scene driving risk data according to a ratio of the virtual scene transmission delay to the accident reaction time when the virtual scene transmission delay is smaller than the accident reaction time, so as to obtain modified driving risk data; and when the virtual scene sending time delay is not less than the accident response time, correcting the virtual scene driving risk data according to the ratio of the accident response time to the virtual scene sending time delay to obtain corrected driving risk data.
11. The apparatus of claim 8, wherein the number of virtual scenes is plural; the data matching module is further used for sorting the corrected driving risk data and the historical accident rate corresponding to each virtual scene of the simulation test vehicle according to the numerical values; constructing a data pair according to the corrected driving risk data and the historical accident rate which have the same sequencing sequence; and when the corrected driving risk number in the data pair is different from the virtual scene corresponding to the historical accident rate, marking the data pair as a matching failure data pair.
12. The apparatus according to claim 11, wherein the validity determination module is further configured to obtain ratio data of the number of matching failure data pairs to the number of data pairs according to the number of data pairs corresponding to the simulation test vehicle and the number of matching failure data pairs; when the ratio data is smaller than the historical accident rate, the matching result is that the corrected driving risk data and the historical accident rate are in a forward relation, and the virtual scene is judged to be effective; and when the ratio data is not less than the historical accident rate, judging that the virtual scene is invalid if the corrected driving risk data and the historical accident rate are not in a forward relation according to the matching result.
13. The apparatus of claim 11, wherein the number of the simulation test vehicles is plural; the validity judging module is further used for collecting the number of the data pairs corresponding to each simulation test vehicle and the number of the matching failure data pairs to obtain the total number of the data pairs and the total number of the matching failure data pairs; determining ratio data of the total amount of the matching failure data pairs and the total amount of the data pairs, and determining an average value of historical accident rates of all simulation test vehicles; when the ratio data is smaller than the average value, the matching result is that the corrected driving risk data and the historical accident rate are in a positive relation, and the virtual scene is judged to be effective; and when the ratio data is not smaller than the average value, judging that the virtual scene is invalid if the matching result indicates that the corrected driving risk data and the historical accident rate are not in a forward relation.
14. The apparatus according to claim 12 or 13, wherein the virtual scenario validity determination apparatus further comprises a data adjustment module, configured to adjust the virtual scenario transmission delay or the accident response time preset by the simulation test vehicle when the virtual scenario is determined to be invalid, so as to obtain a valid virtual scenario.
15. An autopilot system comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, carries out the steps of the method according to one of claims 1 to 7.
16. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011263668.7A CN112256590B (en) | 2020-11-12 | 2020-11-12 | Virtual scene effectiveness judgment method and device and automatic driving system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011263668.7A CN112256590B (en) | 2020-11-12 | 2020-11-12 | Virtual scene effectiveness judgment method and device and automatic driving system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112256590A CN112256590A (en) | 2021-01-22 |
| CN112256590B true CN112256590B (en) | 2022-04-29 |
Family
ID=74265803
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011263668.7A Active CN112256590B (en) | 2020-11-12 | 2020-11-12 | Virtual scene effectiveness judgment method and device and automatic driving system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112256590B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112997128B (en) * | 2021-04-19 | 2022-08-26 | 华为技术有限公司 | Method, device and system for generating automatic driving scene |
| CN113849971B (en) * | 2021-09-16 | 2023-03-28 | 广州文远知行科技有限公司 | Driving system evaluation method and device, computer equipment and storage medium |
| CN116010237B (en) * | 2022-04-18 | 2024-04-05 | 上海滴滴沃芽科技有限公司 | Test method, device, server, medium and product of driving system |
| CN116662474B (en) * | 2023-07-28 | 2023-11-10 | 智道网联科技(北京)有限公司 | High-precision map data processing method, device, equipment and medium |
Family Cites Families (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10671514B2 (en) * | 2016-11-15 | 2020-06-02 | Inrix, Inc. | Vehicle application simulation environment |
| CN108241762A (en) * | 2016-12-26 | 2018-07-03 | 乐视汽车（北京）有限公司 | For the test system and test method of automatic Pilot analogue system |
| US11644834B2 (en) * | 2017-11-10 | 2023-05-09 | Nvidia Corporation | Systems and methods for safe and reliable autonomous vehicles |
| EP3486766A1 (en) * | 2017-11-17 | 2019-05-22 | Steinbeis Interagierende Systeme GmbH | Computer-implemented method of augmenting a simulation model of a physical environment of a vehicle |
| US10713148B2 (en) * | 2018-08-07 | 2020-07-14 | Waymo Llc | Using divergence to conduct log-based simulations |
| US10795804B1 (en) * | 2018-08-21 | 2020-10-06 | Waymo Llc | Collision evaluation for log-based simulations |
| CN109446371A (en) * | 2018-11-09 | 2019-03-08 | 苏州清研精准汽车科技有限公司 | A kind of intelligent automobile emulation testing scene library generating method and test macro and method |
| CN109520744B (en) * | 2018-11-12 | 2020-04-21 | 百度在线网络技术（北京）有限公司 | Driving performance testing method and device for automatic driving vehicle |
| US20200226226A1 (en) * | 2019-01-10 | 2020-07-16 | Uatc, Llc | Autonomous Vehicle Service Simulation |
| CN109993849A (en) * | 2019-03-22 | 2019-07-09 | 山东省科学院自动化研究所 | A kind of automatic Pilot test scene render analog method, apparatus and system |
| CN110263381A (en) * | 2019-05-27 | 2019-09-20 | 南京航空航天大学 | A kind of automatic driving vehicle test emulation scene generating method |
| CN110197027B (en) * | 2019-05-28 | 2023-07-18 | 百度在线网络技术（北京）有限公司 | Automatic driving test method and device, intelligent equipment and server |
| CN110160804B (en) * | 2019-05-31 | 2020-07-31 | 中国科学院深圳先进技术研究院 | Test method, device and system for automatically driving vehicle |
| CN110795818B (en) * | 2019-09-12 | 2022-05-17 | 腾讯科技（深圳）有限公司 | Method and device for determining virtual test scene, electronic equipment and storage medium |
| CN110793784B (en) * | 2019-09-16 | 2021-10-26 | 腾讯科技（深圳）有限公司 | Test method and device for automatic driving vehicle, storage medium and electronic device |
| CN111489588B (en) * | 2020-03-30 | 2024-01-09 | 腾讯科技（深圳）有限公司 | Vehicle driving risk early warning method and device, equipment and storage medium |
-
2020
- 2020-11-12 CN CN202011263668.7A patent/CN112256590B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN112256590A (en) | 2021-01-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112256590B (en) | Virtual scene effectiveness judgment method and device and automatic driving system | |
| CN113642633B (en) | Method, device, equipment and medium for classifying driving scene data | |
| CN108334055B (en) | Method, device and equipment for checking vehicle automatic driving algorithm and storage medium | |
| CN109358526B (en) | Software in-loop testing method of automatic driving tracking algorithm based on simulation | |
| CN111123920A (en) | Method and device for generating automatic driving simulation test scene | |
| CN108508881B (en) | Automatic driving control strategy adjusting method, device, equipment and storage medium | |
| CN112286206B (en) | Automatic driving simulation method, system, equipment, readable storage medium and platform | |
| CN113343461A (en) | Simulation method and device for automatic driving vehicle, electronic equipment and storage medium | |
| CN111492320A (en) | Behavior model of environmental sensor | |
| CN114758504A (en) | Online vehicle overspeed early warning method and system based on filtering correction | |
| CN110751040B (en) | Three-dimensional object detection method and device, electronic equipment and storage medium | |
| CN111947669A (en) | Method for using feature-based positioning maps for vehicles | |
| CN111723672B (en) | Method and device for acquiring video recognition driving track and storage medium | |
| CN113918615A (en) | Simulation-based driving experience data mining model construction method and system | |
| CN112509321A (en) | Unmanned aerial vehicle-based driving control method and system for urban complex traffic situation and readable storage medium | |
| CN116010854B (en) | Abnormality cause determination method, abnormality cause determination device, electronic device and storage medium | |
| CN112629547A (en) | Method and apparatus for creating positioning map | |
| CN116543271A (en) | Method, device, electronic equipment and medium for determining target detection evaluation index | |
| CN115140029A (en) | Safety capability detection method and device for automatic driving automobile | |
| Pan et al. | Metamorphic testing for autonomous driving systems in fog based on quantitative measurement | |
| CN113029155A (en) | Robot automatic navigation method and device, electronic equipment and storage medium | |
| CN111753901A (en) | Data fusion method, device and system and computer equipment | |
| US20230025579A1 (en) | High-definition mapping | |
| CN115063594B (en) | Feature extraction method and device based on automatic driving | |
| Cheng et al. | An Indoor Rapid Testing Platform for Autonomous Vehicles Using Vehicle-in-the-Loop Simulation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |